F rom the end of World War II to the late 1970s, the United States dominated new technology development through massive defense spending and the commanding lead it accumulated during the Cold War. Now the lead is gone, the defense dollars are going, and, increasingly, commercial competitors define the "state of the art." The economic benefits that accompanied technological hegemony-good jobs, wealth, national well being-will be much harder to come by.
The United States can still pioneer technological development at home and exploit innovations that arise abroad-but only if we are willing to make the investments necessary to remain a leading player in global technology. In the past, the private sector has not accomplished that on its own, and it is even less likely to in an age of footloose multinational corporations.
Since World War II and even before the civilian U.S. economy has indirectly benefited from substantial federal funding of science and technology, largely for defense purposes. The record suggests that, except for a handful of highly visible failures where the government tried to overspecify commercial solutions (like the fast-breeder nuclear reactor), publicly supported industries have paid off handsomely in economic growth, jobs, and national prosperity. Some of the "winners" include the nation's leading industries in world markets, like electronics and aerospace. Indeed, according to a 1988 Department of Commerce study, five of the top six fastest growing U.S. industries from 1972 to 1988 were sponsored or sustained, directly or indirectly, by federal investment: computing equipment, semiconductors, optics, imaging technologies, and biological products. (The only exception was lithographic services.)
Yet despite the obvious success of such investment, the debate on public support remains stuck on the same old question: whether the government should be in the game of picking winners and losers. We stand to miss a historic opportunity afforded by the end of the Cold War if we continue debating what is not only necessary but unavoidable. Rather, we should focus on how to reallocate the dollars that become available as defense spending winds down. The goal should be to invest in ways that simultaneously address pressing national problems and foster the commercial technological capabilities that will sustain the American economy into the next century.
WHY HIGH TECHNOLOGY MATTERS
America's technological leadership has dwindled precipitously over twelve long years of ideological opposition to its public support. Studies from sources as diverse as Japan's Ministry of International Trade and Industry, the National Research Council, defense and nondefense agencies of the U.S. government, the European Community, and private-sector business groups reach the same general conclusions: The number of commercial technologies in which the United States has a clear lead over Asia and Europe has dropped sharply during the last decade, a clear reflection of the fact that, as a percentage of gross national product, the U.S. now spends about half as much developing commercial technology as Japan and Germany do.
Even in the aircraft industry, a major stronghold of U.S. high-tech competence, Europe has reached technological parity, and significant advanced production activities are migrating from the U.S. to Asia. This loss of technological position and the post-Cold War cut in public funding have significant implications for the long-run performance of the American economy. Even economists skeptical about public support to commercial technology development admit that technological progress fuels economic growth and that research and development (R&D) fuels technological progress. In fact, at least one-quarter and perhaps as much as one-half of the total growth rate is attributed to advances in technical know-how.
To paraphrase Richard Darman's smirking dismissal of high technology, a dollar's worth of potato chips may indeed be the equivalent of a dollar's worth of silicon chips for the nation's trade balance. But the two are not equivalent for many other determinants of economic well-being such as employment, wages, skill formation, productivity, investment, and research. Indeed, ideological apologists aside, there is emerging a bipartisan consensus that industries can be strategic to economic welfare in one of two ways: by providing a higher return than the same labor, capital, and other inputs could earn elsewhere or by providing spillover benefits for the rest of the economy.1
High-technology industries are strategic in both ways. They fund a disproportionately large amount of industrial R&D nearly 60 percent offering innovations and skills that spill over to the economy as a whole. These spillovers, acknowledged by most economists to provide broad social benefits, are embodied in, for example, basic scientific breakthroughs with myriad applications as well as trained people capable of starting new enterprises. Studies show that as a result of spillovers of all kinds, the social returns to R&D spending far exceed the private returns perhaps by as much as 50 percent to 100 percent above the return to ordinary capital.
High-tech industries also make disproportionate contributions to the nation's creation of high-skill, high-wage employment opportunities. High-tech industries account for only 18 percent of total manufacturing employment but at least one-third of the employment of scientists and engineers. High-tech industries are also high-productivity industries that pay higher compensation than other manufacturing industries. In 1989, the average value added per worker in all high-technology industries was one-third higher than the average for all manufacturing and two-thirds higher if only production workers are included.
THE TECHNOLOGY STAKES
What do computers, supersonic transport, biotechnology, synfuel plants, jet engines and airframes, and the fast-breeder nuclear reactor have in common? All have received, directly or indirectly, intensive public investment and other kinds of public support. With the clarity of hindsight, we can also name the commercial winners and losers.
The winners computers, biotechnology, jet engines, and airframes were each the by-product of public spending for national defense and public health, under that old-fashioned (pre-Reagan) image of government serving the common good. These industries are today leading high-tech sectors of the economy and strong sources of comparative advantage for the nation. More often than not, the public contribution to their success has been discounted, forgotten, or ignored as politically and ideologically inconvenient to the exercise of private power.
The losers supersonic transport, synfuels, and the fast-breeder reactor were each an attempt to fund directly commercial development of a specific product with public money. They were bigtime losers, and the public support that nurtured them is usually very well remembered by opponents of government technology policy. Unlike private investors who focus on the winners, the political debate on government investment obsesses over the losers.
Federal support to new technology crystallized after World War II around national defense, nuclear energy, and, later, space exploration. The spending model was premised on belief that pouring the investments into science at the front end of the development pipelines would produce technology out the other end. Initial applications were developed for the military; later they would spin off into commercial use. In this way, U.S. defense spending promoted the rapid development of a host of military technologies that eventually found widespread success in commercial markets, including jet aircraft and engines, silicon chips, computers and operating systems, complex machine tools, data networks, data compression, optoelectronics, and advanced ceramic and composite materials.
In these cases, government underwrote the relevant basic science research at universities and labs, direct R&D contracts accelerated the development of the technology, and defense procurement at premium prices created a strong initial launch market. A variety of mechanisms, ranging from patent pooling and hardware leasing (machine-tool pools, for example) to loan guarantees for building production facilities, helped to lower market entry costs, to diffuse technology among competitors, and to set the stage for commercial market penetration.
The government provided full-blown support in other sectors, notably for public health (and broadly for science research via the National Science Foundation). Massive government funding for biomedical research and training research scientists followed World War II successes in developing penicillin and other pharmaceuticals. Commercial winners have ranged from treatment regimes, drugs, and medical equipment to biotechnology.
The key to the successful cases seems to have been a development process that met the requirements of the commercial marketplace. Thus when the military pushed silicon-chip design and manufacture toward high reliability, miniaturization, high performance, and low costs, it was creating a trajectory that the commercial computer industry could ride. Similarly, when the Defense Department turned to the scientific community to help define the characteristics for ARPANet, its internal electronics communications system, it was launching a data-networking trajectory that would also meet that community's commercial needs.
In that sense, the U.S. government's direct R&D sponsorship has probably been less important for commercial success than its procurement and indirect support, since the latter policies acted to diffuse the fledgling technologies into widespread use. It is important to note that this strategy of public support was not a simple stepchild of the technological successes of World War II. For example, government support to aeronautics predates the war, beginning in earnest with the creation of NASA's precursor, the National Advisory Commission on Aeronautics, which was a vital source of the R&D and testing during the 1920s and 1930s that led to the modern passenger airliner. Of course, that was in the days when we were still willing to be public risk-takers, before the ideological purity set in when government acted rather than believed it shouldn't.
In fact, some of the grandest and most successful experiments in public support to commercial technology occurred back then. RCA, a consortium among the Navy and several corporations, including General Electric and United Fruit, grew out of Woodrow Wilson's concern about Britain's dominance of radio technology. But arguably the most successful program of public support to commercial innovation is the Agriculture Research and Extension System. Dating from the Morrill Act in 1862, the still-evolving system comprised of land-grant colleges, state experimental stations, and other federal and local instistutions has provided education and training, long-term R&D, and new technologies to America's farms. Although not without controversy (its neglect of organic farming methods, for example), it is still widely credited with a major role in making American agriculture the world's most productive.
W hile such successes are suggestive, there is as much to learn from the failures. The defense-energy-space nexus provides robust examples that range from outright flops like the supersonic transport (SST), synfuel plants, and the fast-breeder reactor, to more complex and ambiguous cases like the development of numerical control for machine tools, the Space Shuttle, or photovoltaics. The Air Force, for example, sponsored the development of numerical control technology for machine tools to build advanced aircraft. But the programming language proved too complex for general commercial use; diffusion was slow and civilian application costly. The resulting development path produced a commercially vulnerable U.S. industry that was squeezed by Japanese competitors from the cheaper end of the market and German firms from the more expensive end.
Similarly, the more visible failures moved down technology trajectories that were commercially unacceptable for a variety of reasons. The commercial airliner market was aiming at short-haul and wide-body planes rather than supersonic speeds. The fast-breeder nuclear reactor and synfuel programs were much more expensive than commercial alternatives, particularly after the oil shocks abated. In each case, developers construed performance objectives too narrowly and failed to explore alternative technological paths. Demonstrations and pilots proceeded despite experimental evidence of failure. In other cases, like photovoltaics, political considerations killed development prematurely.
In short, public support for technology runs into trouble when it proceeds down development paths that diverge from commercial market requirements, particularly when it overspecifies an exotic technical solution in the form of a particular product. (Even in those cases, however, there are likely to be important technical spillovers, especially when the program funds generic research.) Those requirements including manufacturability and customer-defined cost and performance standards should not be hard to build into future programs, particularly if a peer-review model is adapted to include potential commercial producers and users.
MARKET IMPERFECTIONS
Though history shows that public investment in commericial technology can pay off, the question remains: Why can't the private market provide the necessary support? There are two kinds of answers.
First, several so-called imperfections underlie the market's failure to invest adequately. Innovators cannot appropriate all of the returns on their investment. Often, they cannot foresee sufficient returns to justify proposed R&D. Conversely, because of strong intellectual property protection, firms often engage in redundant R&D. The combined effect on the system as a whole is overinvestment in some technologies and underinvestment in others.
Second, innovation is contingent on the decisions and actions of many groups developers, producers, and users and thus unpredictable. Different firms evaluate risks differently, apply different capabilities to their technological effort, receive different signals from their customers, and go down different development paths. That is how Sony and JVC/Philips developed two quite distinct formats, Betamax and VHS, for the VCR, and why there are at least thirty industrial approaches to making flat, high-resolution computer displays.
In essence, technology development is a "path-dependent" process: where you end up depends on where you start and want to go. Technical progress involves insights that materialize only through experience in development, production, and use. Rather than being preordained by scientific logic, then, technology development is contingent.
Contingency means that some firms will go down the wrong path, betting on, for example, dedicated word processors rather than personal computers. Others will start down the right path, misevaluate the potential, and abandon the effort. Consider Xerox's famous Palo Alto Research Center, which pioneered many innovations in desktop computing, including workstations, icons for user interface, and the ubiquitous "mouse" pointing device. Yet other companies, not Xerox, successfully exploited these innovations, just as Japanese firms successfully commercialized the U.S.-developed technologies underlying VCRs, camcorders, and miniature flat displays.
Contingency also means that neither innovators nor the private capital markets that fund them are fully capable of evaluating the risks involved. Some bets will pay off big; some not at all. There will always and must always be winners and losers, but they can only be identified in retrospect.
This is as true for private as well as public investment. For every Macintosh, there are several Lisa's (an early flop from Apple). For every IBM, there are several GE's and Westinghouses whose technological bets on mainframe computers failed to pay off. For every Intel, there are assorted, now-defunct Molectros and Qualidynes. For every winner in a venture portfolio, there are untold losers that get nowhere near the publicity.
There is absolutely no evidence beyond the economist's leap of faith that private investment is any more capable than public investment of separating the winners from the losers before the fact. (The major difference, of course, is that private losers exit the market, while publicly backed losers are held to the higher standard of wasting taxpayers' money.) And as Princeton economist Gene Grossman argues, the magnitude of the potential social gains from public spending are sufficiently large to provide a comfortable margin for error in choosing among technologies to back. Thus picking winners and losers is the wrong metaphor to characterize the socially useful and necessary activity of government in supporting that process. Innovation is a gamble. But through public spending, government is actually placing bets to ensure there will be at least some winners for our collective future.
THE INEVITABILITY OF PUBLIC SUPPORT
There are a range of social needs from national defense to infrastructure provi sion that private investors simply will not effectively support because the risk is too high, the payoff too small, or the market too unsuitable. As it goes about providing for these common needs, the government becomes a significant consumer of technology in its own right. Like any other large customer and lead user, the government must pay to get the technology it needs. Very often that means sponsoring research and procurement that launch new industrial capabilities. As we have seen, such sponsorship and procurement often pay off indirectly in the development of commercially successful technology.
New technologies and industries tend to develop in a particular location Silicon Valley, for example. Such regional patterns are evidence of what economists call "local externalities." These externalities can take several forms, including the availability of a specialized pool of labor, specialized supplier networks, and a common knowledge pool through which firms can learn from each other. Local externalities tend to have a self-reinforcing effect regions or nations that have a strong presence in a particular industry tend to generate the specialized inputs and informational networks that in turn make the industry even more competitive over time.
Although local externalities are not unique to high-technology industries, they have special relevance for them. Recall that technology development paths are strongly characterized by learning. Some such knowledge is footloose, that is, embodied in products, blueprints, or open technical forums like published journals. Other knowledge is, however, developed only in conjunction with development and production. That kind of knowledge accumulates in the specialized local assets like labor pools and supplier networks. It is embodied in them and does not diffuse easily.
When Apple goes to Sony to develop the portable Mac Powerbook, when Compaq goes to Citizen for the LTE notebook, when IBM moves microsystem development out of the U.S. to Japan, these American companies are seeking access to precisely such specialized local Japanese assets in this case, embodying know-how in components and microsystems design and integration. Such local capabilities are the probable basis for product differentiation and new technology generation. They help to attract footloose technological know-how from overseas and to exploit it domestically. In other words, technological progress is intimately bound with local capabilities. Unless these capabilities exist, an economy has no enduring potential for operating at the technological frontier, with all that implies for maintaining national well-being.
N ot surprisingly, other industrial nations have recognized this aspect of technological advance. They know that if they organize their economic systems appropriately and underpin them with an aggressive mix of trade and industrial policies, they stand a fair chance of having the economic benefits of high technology accrue locally rather than abroad. This was the case with Japan's coordinated effort to dominate microelectronics and with Europe's Airbus consortium. Taiwan's proposed stake in McDonnell Douglas in return for transferring 60 percent of subcontracted production activities to Taiwan is only the most visible example of this common practice abroad.
Whether through intention or effect, the local and national concentration of technological benefits means that high-tech competition can take on an inherently beggar-thy-neighbor cast. A bigger national share of global high-tech output can mean a bigger national share of good jobs and greater economic prosperity. That is why technology-intensive industries are the source of growing trade friction between the U.S. and its competitors. As high-tech industries have become steadily more important to the economic performance of the developed countries, those countries commit to maintaining a local high-technology production base by leveraging both policies and structural differences in national economies (for example, the difficulty of Japanese market access).
Absent U.S. government support to domestic technology, national economic systems will interact in ways that tend to redistribute advanced technological activities and their benefits from the U.S. to other countries. We will pay a huge price (measured in loss of high-paying jobs, a contracting tax base, declining local communities, and slower economic growth). We will end up poorer. And it matters not at all whether this outcome is intended or simply the effect of competition between differently organized political economies.
The adjustment problems that follow for the U.S. make government support to commercial technology inevitable. Such support can be reactive later at great cost. Or it can be proactive now, aiming to nurture what we know and extend it in ways that address collective needs. If the government must act, how should it act to maximize the likelihood of a positive payback?
THE PAYOFF
A decade ago, in their book Government and Technological Progress, the economist Richard Nelson and his colleagues provided three recommendations for public support to innovation that still hold.
One is to associate government R&D support with procurement or other well-defined public objectives. A second is to define and fund arenas of nonproprietary research and allow the appropriate scientific community to guide R&D allocation. The third is to develop mechanisms whereby potential users guide the allocation of applied research and development funds.
Just three additional points need to be added to those simple guidelines.
First, any "well-defined public objective" ought to have two characteristics. It should be an area of clear common need for national well-being, and where possible, addressing that need should remove an obvious constraint on future prospects for long-term economic growth.
Second, when funding moves from generic, nonproprietary R&D toward applications, four conditions should prevail: research should pursue alternative technological approaches until the best development path is demonstrated; wherever possible, development must incorporate commercial criteria, such as cost, performance, and manufacturability; within those constraints, leading-edge innovation should take precedence over low-tech solutions to promote high-tech capabilities, which pay higher wages; any resulting technology should first be commercially exploited in the U.S. to ensure that the competencies and production activities accrue locally. (After all, if the government is paying the bill, it is entitled to domestic production, even in an age of multinationals.)
Third, and perhaps most important, public support should actively pursue diffusion of new technologies into widespread commercial use within the domestic economy. Aside from the emphasis on commercial market criteria during application development, there are four ways to accomplish this. One is through incentives like investment tax credits or leasing arrangements tailored to encourage early adopters. A second is through education and training. Many of the industries successfully launched through public sponsorship drew from a labor pool trained in the new technologies through federal support. In effect, a supply of new skills shaped demand to fit its characteristics, rather than the reverse.
A third way to accomplish diffusion is by embedding new technologies in public infrastructure wherever possible, thereby making their benefits universally accessible. A fourth way is to adapt the agriculture extension model to industry. The aim would be to assist the small and medium-sized businesses that underpin local and regional economies in effectively adopting new technologies and methods of production. There is already an elaborate network of state agencies that play this kind of role, often in conjunction with local educational institutions attuned to the needs of local business. Minimal federal funding could help build this network into an effective national industrial extension service and link it to publicly sponsored innovation.
With only marginal effort, these guidelines could be applied to existing federal support of science and technology. But we can go further. Let's imagine redirecting several billion dollars from defense to new commercial technology development. Using existing work on critical technologies as a point of departure, the Office of Science and Technology Policy (or perhaps the National Security Council) would invest one-third of the total on nonproprietary research in the sciences underlying the critical technologies, subject to review by the NSF or the National Academies.
Another one-third would be directed to diffusion, via incentives, training, infrastructure, and extension. This spending would focus on existing technologies like those for flexible manufacturing. This funding could also help move critical technologies out of the lab and into manufacture. For example, demonstration products might be funded or loan guarantees provided for building volume production facilities where private capital markets were known to balk.
The last third would go toward specific nondefense public objectives, where investment is likely to remove constraints on future economic performance. Guided by both technological and commercial market considerations, the resulting technologies should be capable of broad and rapid diffusion into the economy. As serious candidates, consider the following:
The Environment. No matter how much they are opposed by industry, global requirements for ever more stringent environmental cleanliness will not be moderated. But neither are consumers likely to adjust their demand for jobs and goods according to what is environmentally sustainable with today's technologies. These environmental and economic needs can be treated as antithetical, the position President Bush advanced at the Rio Summit in July. Perhaps ultimately they are. But it is surely an appropriate public responsibility to explore whether they can be reconciled.
There are several opportunities that go beyond conventional programs in waste reduction, cleanup, and recycling. The effort should be directed to replacing existing industrial production with technologies that generate no waste or pollution in the first place. If American's manufacturing stock can be rebuilt around clean production technologies, the economic possibilities for the next century are wide open. New U.S. industries to supply the technologies would develop, while established industries that adopted the technologies would have a strong new source of competitive advantage in environmentally sensitive world markets.
Energy and Natural Resources. Almost across the board, American industries are among the least energy- and resource-efficient producers in the industrial world. If there is motivation to develop the appropriate technologies, dramatic resource efficiencies are possible, as Silicon Valley firms discovered when they reduced their water usage during the last six years of drought.
A good beginning would be to study the energy and resource use of major industries like electronics and to evaluate resource-efficient technologies already used abroad. Both tasks are ideally suited to the national laboratories. Once opportunities for improvement are identified, competitive contracts could be offered to develop the necessary technologies, and investment incentives could promote rapid adoption. Again, new technologies to help sustain resources and enable efficient use would spin off entirely new industries while boosting the competitive position of existing ones.
Infrastructure. There is widespread agreement that much of the nation's networks for transportation, power, sewage, water, and communications are eroding and need to be rebuilt. Anyone who has operated a business where the infrastructure is lousy say, the phone networks in Eastern Europe knows the damage it wreaks on efficiency. Providing modern infrastructure would stimulate not only productivity but also innovations, from low-maintenance structural concretes to optical networks, to take advantage of the new infrastructure.
The additional opportunity here would be to fashion new infrastructure that uses emerging critical technologies of the next century new materials, visual systems with flat-panel displays, real-time electronic controls and then to support the development of commercial, domestic capabilities in them through nationwide procurement.
With military spending and the resulting technological and economic benefits it generates declining, these long-term needs are likely to provide the next frontiers of American technological innovation. More than technological preeminence is at stake. Historically high growth rates, competitive wages, and a growing standard of living all depend on regaining leadership in commercial technological innovation. For twelve debilitating years, the government has refused to invest in America's commercial technology position. It's high time to revive public support and set it to the task.